JOURNAL OF CLINICAL MICROBIOLOGY, Aug. 1979, p. 121-127 0095-1137/79/08-0121/07$02.00/0
Vol. 10, No. 2
Fluorescent Antibody Test Kit for Rapid Detection and Identification of Members of the Bacteroides fragilis and Bacteroides melaninogenicus Groups in Clinical Specimens JOSEPH W. HOLLAND,* LARRY R. STAUFFER,t AND WILLIAM A. ALTEMEIER Research Surgical Bacteriology Laboratory, Department of Surgery, University of Cincinnati, College of Medicine, Cincinnati, Ohio 45267 Received for publication 10 May 1979
The Fluoretec fluorescent antibody test kit (Pfizer Inc., New York, N.Y.), developed for the rapid detection of members of the Bacteroides fragilis and B. melaninogenicus groups, was evaluated by testing 58 stock cultures and 76 clinical specimens. The test reagents detected 100% of 40 B. fragilis and B. thetaiotaomicron stock culture strains, although only 22% of 18 B. vulgatus, B. distasonis, and B. ovatus strains showed positive fluorescence. The 76 clinical specimens were evaluated by examining fluorescent antibody-stained smears of 49 specimens of purulent material and smears of 27 blood cultures which were positive for gram-negative bacilli by Gram stain or subculture. The Fluoretec reagent detected members of the B. fragilis group in 28 (97%) of the 29 specimens of purulent material and all (100%) of the 16 blood cultures in which these anaerobes were demonstrated by culture. Overall, the Fluoretec reagent detected members of the B. fragilis group in 44 (98%) of the 45 clinical specimens which were shown by culture to contain these anaerobes. Two of the 76 clinical specimens gave positive fluorescence for members of the B. fragilis group but failed to yield these organisms by culture. Members of the B. melaninogenicus group were detected by culture in 15 specimens and in each case their presence was demonstrated by the Fluoretec reagent. No members of the B. melaninogenicus group were isolated from five clinical specimens that gave positive fluorescence with the B.
melaninogenicus reagent. The involvement of anaerobic bacteria in a The need exists for a means of detecting and wide variety of infections is well documented. identifying the clinically significant anaerobes The anaerobes most frequently isolated from soon enough to influence patient care. Griffin (6) these infections, the Bacteroides, are also fre- reported the feasibility of using fluorescent anquently resistant to the commonly used antimi- tibody (FA) for the rapid identification of species crobial agents. Over a 9-year period, bacteroides of Bacteroides, Fusobacterium, and Sphaerowere isolated from nearly 70% of the anaerobe- phorus in clinical specimens. Studies by Garcia positive specimens cultured in this laboratory et al. (5) and Fales and Teresa (2) also supported (7). Members of the B. fragilis group were found the feasibility of FA for the rapid detection of S. in 52% of the anaerobe-positive specimens and necrophorus (F. necrophorum). Lambe (9) apaccounted for a greater percentage of the anaer- plied the FA technique for the classification of obic isolates (24%) than any other species (7). B. melaninogenicus. Stauffer et al. (11) develMembers of the B. melaninogenicus group were oped and evaluated an indirect FA technique for the third most frequent isolates and accounted the rapid detection and identification of Bactefor 7% of the anaerobes recovered. roides and Fusobacterium. Recently, a commerThere are several conventional and microtest cial FA test kit (Fluoretec) has been developed systems available for identification of anaerobic for the detection of members of the B. fragilis bacteria. These systems involve biochemical test and the B. melaninogenicus groups in clinical reactions and therefore require isolation and specimens. This report describes a clinical evalpure culture growth which may delay identifi- uation of the Fluoretec FA test kit. cation by 2 to 7 days. Results are often obtained MATERIALS AND METHODS too late to benefit the clinician or the patient and become of academic interest only. Bacterial strains. Fifty-six of the 58 Bacteroides t Present address: Clinical Microbiology Laboratory, Cin- strains used in this study were isolated from a variety cinnati General Hospital, Cincinnati, OH 45267. of clinical specimens in the Research Surgical Bacte121
122
HOLLAND, STAUFFER, AND ALTEMEIER
riology Laboratory and stored at -60°C in broth culture. The strains were identified as previously described (7). One strain of B. vulgatus, WAL 1887, was obtained from V. L. Sutter, Wadsworth Hospital Center, Los Angeles, Calif. The single strain of B. ovatus was American Type Culture Collection strain ATCC 8483. Frozen stock cultures were prepared for FA testing by picking isolated 48-h colonies from plates of blood agar base (BBL Microbiology Systems, Cockeysville, Md.) containing 4% defibrinated rabbit blood, hemin (5 ug/ml, final concentration), and menadione (0.5 pg/ml, final concentration) (RHM plates). The colonies were suspended in saline to the turbidity of one-half of a McFarland no. 1 tube, and the bacterial cell suspension was swabbed onto clean glass slides and air dried. In this report, the term B. fragilis group (BFG) refers to members of the five previously recognized subspecies of B. fragilis. The designation B. fragilis is reserved for the species B. fragilis (1). The term B. melaninogenicus group (BMG) refers to the three previously recognized subspecies of B. melaninogenicus and includes those organisms presently classified as B. asaccharolyticus, B. melaninogenicus subsp. melaninogenicus, and B. melaninogenicus subsp. intermedius (4). B. melaninogenicus subsp. intermedius and subsp. melaninogenicus are not differentiated here and are listed collectively as B. melaninogenicus. Clinical specimens. The 76 clinical specimens tested were obtained from patients at the Cincinnati General and the Christian R. Holmes Hospitals, Cincinnati, Ohio. Specimens were collected and transported as previously described (7). Specimens of blood were inoculated directly into aerobic and anaerobic blood culture bottles (B6053 and B6061, GIBCO Diagnostics, The Mogul Corp., Chagrin Falls, Ohio) and subcultured aerobically and anaerobically on RHM or sheep blood agar (SHM) plates (blood agar base [BBL] containing 4% defibrinated sheep blood, 5 pg of hemin per ml, and 0.5 pg of menadione per ml). Specimens of purulent material, upon reaching the laboratory, were immediately transferred to an anaerobic glove box (Coy Manufacturing Co., Ann Arbor, Mich.) containing 85% nitrogen, 10% hydrogen, and 5% carbon dioxide. The minimum combination of isolation media used was thioglycolate medium (BBL Microbiology Systems) and SHM or RHM plates incubated aerobically and anaerobically. These media were often supplemented with eosin-methylene blue agar (BBL Microbiology Systems) (incubated aerobically); SHM or RHM containing kanamycin (100 ,ug/ml, final concentration), kanamycin-vancomycin blood agar (GIBCO, P-2145) (incubated anaerobically), or both; and phenylethyl alcohol blood agar (BBL Microbiology Systems) or colistin, nalidixic acid blood agar (GIBCO, P2143) (incubated aerobically and anaerobically). Anaerobic strains isolated from clinical specimens were identified using 0.5-ml volumes of biochemical test media and gas-liquid chromatographic analysis of short-chain fatty acids as previously described (8). Microscopy. A Leitz Wetzler Ortholux microscope with an Osram HBO 200-W mercury lamp was used for this study. Filters included a heat-absorbing BG 38 and an FITC interference-type filter having a maximum transmission range of 450 to 490 nm.
J. CLIN. MICROBIOL.
FA test kit. The FA test kit (Fluoretec, Pfizer Inc., New York, N.Y.) that we evaluated contained a Formalin-fixed cell preparation of the BFG and the BMG for use as positive controls, a prestain of rhodamineB-labeled normal human gamma globulin, a polyvalent fluorescein isothiocyanate-labeled antibody pool directed against members of the BFG, a polyvalent fluorescein-labeled antibody pool directed against members of the BMG, and buffered glycerol mounting fluid. The development of the Fluoretec reagents will be described in a forthcoming publication (B. Z. Horvath and G. E. Lowke, Pfizer Inc., personal commu-
nication). Procedure. As soon as growth of gram-negative bacteria was detected in blood culture medium by Gram stain or subculture, samples were aspirated, spread on glass slides, and air dried for FA staining. Specimens of purulent material were removed from the glove box immediately after initial plating and smeared and air dried. The staining procedure used during the present studies was as follows. Air-dried smears of the specimen or control cells were gently heat fixed by two to three passages through a flame, or as an alternative were methanol fixed for 10 min. Fixed smears were reacted with rhodamine prestain for 15 min in a moist chamber at room temperature. The prestain was removed gently by blotting, and the smears were reacted with fluorescein-labeled antibodies to members of the BFG or the BMG. After 15 to 20 min in a moist chamber at room temperature, the smears were rinsed by a 5-min immersion in 0.01 M phosphate-buffered saline (pH 7.4) followed by a gentle flush with distilled water. The rinsed smears were air dried and mounted in buffered glycerol. Each time specimens were stained, positive and negative controls were performed by reacting BFG and BMG control cells with the labeled BFG and BMG antibody pools. Controls and specimens were considered positive if bacterial cells showed peripheral fluorescent staining. FA test results were compared on a blind basis with the culture results obtained on each specimen. If clinical specimens showed positive FA results for the BFG, the BMG, or both, but the corresponding organism was not isolated (false positives), pure culture growth of each strain isolated from the clinical specimen was tested to determine whether other bacteria were reacting with the antibody pools. Members of the BFG and the BMG which were isolated from specimens that failed to show fluorescence (false negatives) were also tested in pure culture against the antibody pools.
RESULTS The results obtained by staining 58 stock culture strains of the BFG with the Fluoretec BFG polyvalent antibody reagent are shown in Table 1. Seventy-six percent of the 58 strains showed positive fluorescence, 5% gave borderline reactions, and 19% failed to react. Each of the 40 B. fragilis and B. thetaiotaomicron strains, tested on a blind basis, showed positive fluorescence. Of the 10 B. vulgatus strains tested, 4 gave positive fluorescence, 2 gave borderline reac-
EVALUATION OF FLUORETEC FA SYSTEM
VOL. 10, 1979
TABLE 1. Reaction of 58 stock culture strains of the B. fragilis group with the Fluoretec BFG reagent. Fluorescence Bacteria
B. fragilis B. thetaiotaomicron B. vulgatus B. distasonis B. ovatus
No.
tested
Positive
Borderline
Negative
28 12 10 7 1
28 12 4 0 0
0 0 2 1 0
0 0 4 6 1
tions, and 4 failed to fluorescence. The 7 B. distasonis strains and the single B. ovatus strain failed to gave positive reactions. Smears of 27 blood cultures and 49 specimens of purulent material were stained by using the Fluoretec reagents. Of these 76 specimens, 45 were positive for members of the BFG by culture. Forty-four (98%) of the 45 culture-positive specimens were positive when stained with the Fluoretec BFG reagent. The 27 blood cultures showed complete agreement between FA and culture results (Table 2). The Fluoretec BFG reagent was positive for members of the BFG in 16 of the blood cultures, and in each case a member of the BFG was the only organism isolated. Of the remaining 11 blood cultures, which were negative for the BFG by FA and culture, one yielded a Clostridium and 10 yielded aerobic gram-negative bacilli. The one false-negative and the two false-positive results which were obtained from the 76 clinical specimens were from specimens of purulent material (Table 3). The single false-negative specimen (R78091) failed to give positive fluorescence even though B. fragilis was isolated. The direct smear of this specimen showed an abundance of brightly fluorescing material which appeared to be debris, but no bacteria with typical morphology and peripheral staining. The B. fragilis isolated from this specimen gave negative to weak overall fluorescence with an occasional cell staining more intensely. Two specimens of purulent material (R77008 and X78061) gave positive fluorescence for members of the BFG, but failed to yield these organisms by culture. One anaerobic gram-negative bacillus isolated from specimen X78061 was not able to be subcultured and therefore could not be identified or tested in pure culture with the FA reagents. The only other anaerobe isolated from this specimen failed to fluoresce with the BFG reagent as did the isolates from specimen R77008. Members of the BMG were detected by culture in 15 of the 76 clinical specimens. The Fluoretec BMG reagent detected members of the BMG in each of the 15 specimens (100%).
123
TABLE 2. Culture and FA results obtained from 27 blood cultures No. of cultures
16 1
FA
Culture
Isolates
BMG
BFG
FA
Culture
Aerobic
Anaerobic
-
None None
BFG Clostridium
+
+
-
-
-
-
sp. GNR' 10 a GNR, Gram-negative rod.
None
Five specimens were FA positive for the BMG, although these anaerobes could not be isolated. All aerobic and anaerobic isolates from each of the five specimens (X77081, X77082, R77084, R77098, and X78064) which demonstrated a false-positive reaction were tested individually with the BMG reagent. None of the organisms isolated from these clinical specimens gave positive results. The BFG and BMG reagents did not appear to react with aerobic or other anaerobic bacteria in clinical specimens; however, the BMG staining reagent did occasionally react weakly with Formalin-fixed BFG control cells. DISCUSSION The results obtained by testing 58 stock culture strains of the BFG show that the Fluoretec BFG reagent detected 100% of the 40 B. fragilis and B. thetaiotaomicron strains, but detected only 22% of 18 B. vulgatus, B. distasonis, and B. ovatus strains. The low percentage of detection of the latter three species did not appear to be a problem when testing direct clinical smears since there were no false-negative results due to the failure of these species to fluoresce. However, there were only two specimens from which these species were isolated in the absence of B. fragilis or B. thetaiotaomicron. The B. distasonis and the B. vulgatus strains present in these specimens did fluoresce, but two specimens is an insufficient number on which to evaluate the effectiveness of the BFG reagent for detection of B. vulgatus, B. distasonis, or B. ovatus in direct clinical smears. It should also be noted that results based on the data obtained with stock cultures (Table 1) may be misleading due to the small number of B. vulgatus, B. distasonis, and B. ovatus stock cultures tested and since some strains, after storage at -60°C, may not fluoresce with the same intensity as strains found in fresh clinical specimens. Although studies using stock cultures indicate inadequate coverage of these members of the BFG, the need for more complete FA coverage of B. vulgatus, B. distasonis, and B. ovatus may be minimized by the fact that these species are
124
J. CLIN. MICROBIOL.
HOLLAND, STAUFFER, AND ALTEMEIER
TABLE 3. Culture and FA results obtained from 49 clinical specimens ofpurulent material Culture results'
FA results
Source and culture no.
Abdominal fluid (R75086) Brain abscess (R77008) Liver abscess (R77023)
(R77024) Peritoneal fluid (X77062) Subphrenic abscess (R77064)
BFG
BMG
+
+
Aerobic
None B. asaccharolyticus Bacteroides sp. GPC
GNR GNR None GNR GPC
None None B. fragilis B. fragilis Bacteroides sp. F. nucleatum B. fragilis B. vulgatus B. thetaiotaomicron GPC None None None B. fragilis B. asaccharolyticus GPC (3) B. fragilis B. thetaiotaomicron B. distasonis B. vulgatus GPR None B. thetaiotaomicron GPC B. melaninogenicus GPR, GPC (2) Fusobacterium sp. Bacteroides sp. GPR GPC
Granulation tissue (R77066)
GNR (2) GPC GPR (2)
Subcutaneous abscess (R77073) Leg ulcer (R77077) Sinus tract (R77079) Gangrenous tissue (R77080)
None None GPR GNR (2) GPC (3)
Colon suture site (X77081)
GNR
Sinus tract (X77082) Osteomyelitis (R77084)
None GNR (3)
Pus (X77096)
None
Pleural fluid (X77097)
None
Fistula tract (R77098)
Colostomy drainage (X77099)
GNR GPC GNR (3) GPC
Osteomyelitis (X77101)
GNR GPC
Retroperitoneal abscess (R78001) Abdominal abscess
GNR
(R78002) (X78003)
GNR (2) GPC (2) GNR (2) GPC (2)
Pelvic abscess (X78004)
GPC
Buttock abscess (X78006)
GPC (2) GNR Yeast GPC (2) GNR (2) GNR (3) GNR
Jejunal wall (X78007) Peritoneal fluid (X78008) Subphrenic abscess (X78009)
Anaerobic
GNR None
B. fragilis B. melaninogenicus GPR, GPC (2) B. fragilis B. melaninogenicus GNR, GPC B. fragilis B. fragilis B. thetaiotaomicron B. thetaiotaomicron B. vulgatus GPC GPR B. fragilis GPC None GPR (2) GPR (2) B. fragilis
VOL. 10, 1979
EVALUATION OF FLUORETEC FA SYSTEM
125
TABLE 3-Continued Culture resultsa
FA results Source and culture no. BFG
BMG
Pilonidal abscess (X78011)
+
+
None
Sinus tract (X78012) Abdominal wound (R78018)
+
+
None GNR (2) GPC (2)
Pus, chest wall (R78021)
+
+
GPC (2) GNR
Pus (X78023)
+
+
GNR (2)
(X78024)
+
+
GNR
Buttock abscess (X78057)
+
+
ND
Bartholin abscess (X78061)
+
+
ND
Pilonidal cyst (R78062)
+
-
GPC
Fistula tract (R78063)
+
-
Abdominal wound (X78064) (X78067)
GNR GPC (4)
+ +
+ +
GNR GNR (2)
Lesion, groin (X78068)
+
+
GNR (2) GPC (2)
Amniotic fluid (X78069) Hip abscess (X78079) Abdominal wound (X78080) Facial abscess (X78086)
-
+
None GPC None None
+ + +
Aerobic
Anaerobic
B. fragilis B. asaccharolyticus GPR, GPC (3) None B. fragilis B. thetaiotaomicron B. distasonis B. melaninogenicus GPR (2), GPC (2) B. fragilis B. melaninogenicus GPC, GPR (2) B. fragilis B. asaccharolyticus GPC B. fragilis B. asaccharolyticus B. vulgatus B. melaninogenicus B. asaccharolyticus GNR B. thetaiotaomicron Bacteroides sp. B. fragilis GPC B. fragilis B. fragilis B. melaninogenicus Bacteroides sp. B. fragilis B. thetaiotaomicron B. melaninogenicus Bacteroides sp. None B. fragilis B. fragilis B. fragilis B. melaninogenicus GPC (3) None
Peritoneal exudate (X78088) None Sinus tract (R78090) GPC None (R78091) None B. fragilis Osteomyelitis (M818) GPC GPR Thoracentesis (M827) + None B. thetaiotaomicron a GNR, Gram-negative rod; GPR, gram-positive rod; GPC, gram-positive coccus; ND, not done. Numbers in parentheses indicate the number of isolates.
isolated from clinical specimens far less frequently than B. fragilis and B. thetaiotaomicron (7, 8), and when found, they are usually accompanied by one or both of these species. The Fluoretec BFG reagent was 100% accurate for detecting B. fragilis, B. thetaiotaomicron, and B. distasonis strains in blood cultures which contained gram-negative bacilli. The
aerobic gram-negative bacilli present in 10 of the blood cultures and the Clostridium species isolated from one specimen did not fluoresce, thus there were no false-positive reactions resulting from these bacteria. Based on this limited number of 27 specimens, it appears that the BFG reagent is a reliable tool for the detection of members of the BFG in blood cultures.
126
HOLLAND, STAUFFER, AND ALTEMEIER
There were no members of the BMG detected in the 27 blood cultures by either the BMG reagent or by culture; therefore, it was not possible to determine from the present study whether these anaerobes could be detected in blood cultures. Based on our results with specimens of purulent material (Table 3), there is reason to believe that a high percentage would be detected; however, the reported incidence of these anaerobes in blood cultures (3, 10) is probably not high enough to justify routine testing for the BMG. With direct smears of purulent material, 98% of the specimens which were positive for members of the BFG by culture were also positive with the Fluoretec BFG reagent. These results are similar to those obtained from blood cultures; however, direct smears of some purulent materials were more difficult to evaluate than blood culture specimens because they contained greater amounts of debris and fewer bacterial cells with typical morphology. These factors probably influenced the evaluation of specimen R78091 (Table 3). In addition, blood cultures were not stained unless a gram-negative bacillus was detected by Gram stain or subculture. Therefore, blood culture specimens were less likely to yield false-positive results due to factors such as prior antibiotic treatment and inadequate care in collection, transportation, and cultivation which could influence the results obtained from specimens that were stained before the demonstration of viable bacteria. These factors may have been responsible for the falsepositive results obtained from specimens R77008 and X78061 (Table 3). False-positive results could also have resulted from the failure to recognize and pick all of the different colony types present on the original isolation plates. Five clinical specimens were positive for members of the BMG when stained with the Fluoretec BMG reagent, yet these bacteria were not detected by culture. This could have resulted from the reasons previously discussed with regard to the BFG. Occasionally, a small percentage of Formalin-fixed BFG control cells fluoresced weakly when reacted with the BMG reagent. This was probably not the cause of the five false-positive reactions since one specimen yielded no growth and the isolates from the remaining four specimens did not fluoresce with the BMG reagent. During this study, the Fluoretec reagents were reliable tools for detecting members of the BFG and the BMG in clinical specimens. The real advantage of using the FA procedure is that members of the BFG and the BMG can be detected in purulent material within 1 to 2 h after the specimen is obtained and in blood
J. CLIN. MICROBIOL.
cultures as soon as or possibly before growth is detected. Also, since bacterial viability is not required, detection may be possible even if the specimen is improperly handled during collection and transportation or if the specimen is obtained after the administration of certain antibiotics. The FA procedure should improve patient care by reducing by several days the time required to presumptively detect the BFG and the BMG, thus allowing the early initiation of
appropriate therapy. As with any FA procedure, the interpretation of results will depend a great deal on the experience and judgment of the individual performing the test, the type and condition of equipment, the proper use of test and control reagents,
and the criteria used to determine what constitutes a positive and a negative result. For this study, clinical specimens were judged to be positive if bacterial cells showed peripheral fluorescence and negative if peripheral fluorescence was not demonstrated. Most specimens gave clear-cut results; however, some gave borderline reactions which were difficult to evaluate. Several factors were found to influence the amount of fluorescence detected. The intensity of fluorescent staining could frequently be increased by extending to 30 min the reaction period of the FA reagent with the specimen. With some specimens, fluorescence was difficult to detect because cells did not adhere sufficiently to slides after heat fixation and were lost during the rinsing and blotting steps. In our experience, a 10-min methanol fixation of airdried slides resulted in greater cell adherence. The type of excitation filter used also affected the degree of staining. Reactions which were questionable with a BG 12 filter were frequently positive with an FITC interference type filter. The proper alignment and use of the microscope are other factors to be considered. It is recommended that potential users of the FA technique be given training in fluorescent microscopy which would include the proper use and care of equipment, the proper use of the test and control reagents, and the opportunity to evaluate known specimens. ACKNOWLEDGMENTS We thank H. S. Bjornson for a critical review of the manuscript. This work was supported in part by Public Health Service grant 5-P50-GM-15428 from the National Institute of General Medical Sciences and in part by Pfizer Inc., New York, N.Y.
LITERATURE CITED 1. Cato, E. P., and J. L. Johnson. 1976. Reinstatement of species rank for Bacteroides fragilis, B. ovatus, B. distasonis, B. thetaiotaomicron, and B. tulgatus: designation of neotype strains for Bacteroides fragilis (Veillon and Zuber) Castellani and Chalmers and Bac-
VOL. 10, 1979
EVALUATION OF FLUORETEC FA SYSTEM
teroides thetaiotaomicron (Distaso) Castellani and Chalmers. Int. J. Syst. Bacteriol. 26:230-237. 2. Fales, W. H., and G. W. Teresa. 1972. Fluorescent antibody technique for identifying isolates of Sphaerophorus necrophorus of bovine hepatic abscess origin. Am. J. Vet. Res. 33:2323-2329. 3. Felner, J. M., and V. R. Dowell, Jr. 1971. "Bacteroides" bacteremia. Am. J. Med. 50:787-796. 4. Finegold, S. M., and E. A. Barnes. 1977. Report of the ICSB taxonomic subcommittee on gram-negative anaerobic rods: proposal that the saccharolytic and asaccharolytic strains at present classified in the species Bacteroides melaninogenicus (Oliver and Wherry) be reclassified in two species as Bacteroides melaninogenicus and Bacteroides asaccharolyticus. Int. J. Syst. Bacteriol. 27:388-391. 5. Garcia, M. M., D. H. Neil, and K. A. McKay. 1971. Application of immunofluorescence to studies on the ecology of Sphaerophorus necrophorus. Appl. Microbiol. 21:809-814. 6. Griffin, M. H. 1970. Fluorescent antibody techniques in
127
the identification of the gram-negative nonsporeforming anaerobes. Health Lab. Sci. 7:78-83. 7. Holland, J. W., E. 0. Hill, and W. A. Altemeier. 1977. Numbers and types of anaerobic bacteria isolated from clinical specimens since 1960. J. Clin. Microbiol. 5:2025.
8. Holland, J. W., S. M. Gagnet, S. A. Lewis, and L. R. Stauffer. 1977. Clinical evaluation of a simple, rapid procedure for the presumptive identification of anaerobic bacteria. J. Clin. Microbiol. 5:416-426. 9. Lambe, D. W., Jr. 1974. Determination of Bacteroides melaninogenicus serogroups by fluorescent antibody staining. Appl. Microbiol. 28:561-567. 10. Martin, W. J. 1974. Isolation and identification of anaerobic bacteria in the clinical laboratory: a 2-year experience. Mayo Clin. Proc. 49:300-308. 11. Stauffer, L. R., E. 0. Hill, J. W. Holland, and W. A. Altemeier. 1975. Indirect fluorescent antibody procedure for the rapid detection and identification of Bacteroides and Fusobacterium in clinical specimens. J. Clin. Microbiol. 2:337-344.
Vol. 10, No. 2
Fluorescent Antibody Test Kit for Rapid Detection and Identification of Members of the Bacteroides fragilis and Bacteroides melaninogenicus Groups in Clinical Specimens JOSEPH W. HOLLAND,* LARRY R. STAUFFER,t AND WILLIAM A. ALTEMEIER Research Surgical Bacteriology Laboratory, Department of Surgery, University of Cincinnati, College of Medicine, Cincinnati, Ohio 45267 Received for publication 10 May 1979
The Fluoretec fluorescent antibody test kit (Pfizer Inc., New York, N.Y.), developed for the rapid detection of members of the Bacteroides fragilis and B. melaninogenicus groups, was evaluated by testing 58 stock cultures and 76 clinical specimens. The test reagents detected 100% of 40 B. fragilis and B. thetaiotaomicron stock culture strains, although only 22% of 18 B. vulgatus, B. distasonis, and B. ovatus strains showed positive fluorescence. The 76 clinical specimens were evaluated by examining fluorescent antibody-stained smears of 49 specimens of purulent material and smears of 27 blood cultures which were positive for gram-negative bacilli by Gram stain or subculture. The Fluoretec reagent detected members of the B. fragilis group in 28 (97%) of the 29 specimens of purulent material and all (100%) of the 16 blood cultures in which these anaerobes were demonstrated by culture. Overall, the Fluoretec reagent detected members of the B. fragilis group in 44 (98%) of the 45 clinical specimens which were shown by culture to contain these anaerobes. Two of the 76 clinical specimens gave positive fluorescence for members of the B. fragilis group but failed to yield these organisms by culture. Members of the B. melaninogenicus group were detected by culture in 15 specimens and in each case their presence was demonstrated by the Fluoretec reagent. No members of the B. melaninogenicus group were isolated from five clinical specimens that gave positive fluorescence with the B.
melaninogenicus reagent. The involvement of anaerobic bacteria in a The need exists for a means of detecting and wide variety of infections is well documented. identifying the clinically significant anaerobes The anaerobes most frequently isolated from soon enough to influence patient care. Griffin (6) these infections, the Bacteroides, are also fre- reported the feasibility of using fluorescent anquently resistant to the commonly used antimi- tibody (FA) for the rapid identification of species crobial agents. Over a 9-year period, bacteroides of Bacteroides, Fusobacterium, and Sphaerowere isolated from nearly 70% of the anaerobe- phorus in clinical specimens. Studies by Garcia positive specimens cultured in this laboratory et al. (5) and Fales and Teresa (2) also supported (7). Members of the B. fragilis group were found the feasibility of FA for the rapid detection of S. in 52% of the anaerobe-positive specimens and necrophorus (F. necrophorum). Lambe (9) apaccounted for a greater percentage of the anaer- plied the FA technique for the classification of obic isolates (24%) than any other species (7). B. melaninogenicus. Stauffer et al. (11) develMembers of the B. melaninogenicus group were oped and evaluated an indirect FA technique for the third most frequent isolates and accounted the rapid detection and identification of Bactefor 7% of the anaerobes recovered. roides and Fusobacterium. Recently, a commerThere are several conventional and microtest cial FA test kit (Fluoretec) has been developed systems available for identification of anaerobic for the detection of members of the B. fragilis bacteria. These systems involve biochemical test and the B. melaninogenicus groups in clinical reactions and therefore require isolation and specimens. This report describes a clinical evalpure culture growth which may delay identifi- uation of the Fluoretec FA test kit. cation by 2 to 7 days. Results are often obtained MATERIALS AND METHODS too late to benefit the clinician or the patient and become of academic interest only. Bacterial strains. Fifty-six of the 58 Bacteroides t Present address: Clinical Microbiology Laboratory, Cin- strains used in this study were isolated from a variety cinnati General Hospital, Cincinnati, OH 45267. of clinical specimens in the Research Surgical Bacte121
122
HOLLAND, STAUFFER, AND ALTEMEIER
riology Laboratory and stored at -60°C in broth culture. The strains were identified as previously described (7). One strain of B. vulgatus, WAL 1887, was obtained from V. L. Sutter, Wadsworth Hospital Center, Los Angeles, Calif. The single strain of B. ovatus was American Type Culture Collection strain ATCC 8483. Frozen stock cultures were prepared for FA testing by picking isolated 48-h colonies from plates of blood agar base (BBL Microbiology Systems, Cockeysville, Md.) containing 4% defibrinated rabbit blood, hemin (5 ug/ml, final concentration), and menadione (0.5 pg/ml, final concentration) (RHM plates). The colonies were suspended in saline to the turbidity of one-half of a McFarland no. 1 tube, and the bacterial cell suspension was swabbed onto clean glass slides and air dried. In this report, the term B. fragilis group (BFG) refers to members of the five previously recognized subspecies of B. fragilis. The designation B. fragilis is reserved for the species B. fragilis (1). The term B. melaninogenicus group (BMG) refers to the three previously recognized subspecies of B. melaninogenicus and includes those organisms presently classified as B. asaccharolyticus, B. melaninogenicus subsp. melaninogenicus, and B. melaninogenicus subsp. intermedius (4). B. melaninogenicus subsp. intermedius and subsp. melaninogenicus are not differentiated here and are listed collectively as B. melaninogenicus. Clinical specimens. The 76 clinical specimens tested were obtained from patients at the Cincinnati General and the Christian R. Holmes Hospitals, Cincinnati, Ohio. Specimens were collected and transported as previously described (7). Specimens of blood were inoculated directly into aerobic and anaerobic blood culture bottles (B6053 and B6061, GIBCO Diagnostics, The Mogul Corp., Chagrin Falls, Ohio) and subcultured aerobically and anaerobically on RHM or sheep blood agar (SHM) plates (blood agar base [BBL] containing 4% defibrinated sheep blood, 5 pg of hemin per ml, and 0.5 pg of menadione per ml). Specimens of purulent material, upon reaching the laboratory, were immediately transferred to an anaerobic glove box (Coy Manufacturing Co., Ann Arbor, Mich.) containing 85% nitrogen, 10% hydrogen, and 5% carbon dioxide. The minimum combination of isolation media used was thioglycolate medium (BBL Microbiology Systems) and SHM or RHM plates incubated aerobically and anaerobically. These media were often supplemented with eosin-methylene blue agar (BBL Microbiology Systems) (incubated aerobically); SHM or RHM containing kanamycin (100 ,ug/ml, final concentration), kanamycin-vancomycin blood agar (GIBCO, P-2145) (incubated anaerobically), or both; and phenylethyl alcohol blood agar (BBL Microbiology Systems) or colistin, nalidixic acid blood agar (GIBCO, P2143) (incubated aerobically and anaerobically). Anaerobic strains isolated from clinical specimens were identified using 0.5-ml volumes of biochemical test media and gas-liquid chromatographic analysis of short-chain fatty acids as previously described (8). Microscopy. A Leitz Wetzler Ortholux microscope with an Osram HBO 200-W mercury lamp was used for this study. Filters included a heat-absorbing BG 38 and an FITC interference-type filter having a maximum transmission range of 450 to 490 nm.
J. CLIN. MICROBIOL.
FA test kit. The FA test kit (Fluoretec, Pfizer Inc., New York, N.Y.) that we evaluated contained a Formalin-fixed cell preparation of the BFG and the BMG for use as positive controls, a prestain of rhodamineB-labeled normal human gamma globulin, a polyvalent fluorescein isothiocyanate-labeled antibody pool directed against members of the BFG, a polyvalent fluorescein-labeled antibody pool directed against members of the BMG, and buffered glycerol mounting fluid. The development of the Fluoretec reagents will be described in a forthcoming publication (B. Z. Horvath and G. E. Lowke, Pfizer Inc., personal commu-
nication). Procedure. As soon as growth of gram-negative bacteria was detected in blood culture medium by Gram stain or subculture, samples were aspirated, spread on glass slides, and air dried for FA staining. Specimens of purulent material were removed from the glove box immediately after initial plating and smeared and air dried. The staining procedure used during the present studies was as follows. Air-dried smears of the specimen or control cells were gently heat fixed by two to three passages through a flame, or as an alternative were methanol fixed for 10 min. Fixed smears were reacted with rhodamine prestain for 15 min in a moist chamber at room temperature. The prestain was removed gently by blotting, and the smears were reacted with fluorescein-labeled antibodies to members of the BFG or the BMG. After 15 to 20 min in a moist chamber at room temperature, the smears were rinsed by a 5-min immersion in 0.01 M phosphate-buffered saline (pH 7.4) followed by a gentle flush with distilled water. The rinsed smears were air dried and mounted in buffered glycerol. Each time specimens were stained, positive and negative controls were performed by reacting BFG and BMG control cells with the labeled BFG and BMG antibody pools. Controls and specimens were considered positive if bacterial cells showed peripheral fluorescent staining. FA test results were compared on a blind basis with the culture results obtained on each specimen. If clinical specimens showed positive FA results for the BFG, the BMG, or both, but the corresponding organism was not isolated (false positives), pure culture growth of each strain isolated from the clinical specimen was tested to determine whether other bacteria were reacting with the antibody pools. Members of the BFG and the BMG which were isolated from specimens that failed to show fluorescence (false negatives) were also tested in pure culture against the antibody pools.
RESULTS The results obtained by staining 58 stock culture strains of the BFG with the Fluoretec BFG polyvalent antibody reagent are shown in Table 1. Seventy-six percent of the 58 strains showed positive fluorescence, 5% gave borderline reactions, and 19% failed to react. Each of the 40 B. fragilis and B. thetaiotaomicron strains, tested on a blind basis, showed positive fluorescence. Of the 10 B. vulgatus strains tested, 4 gave positive fluorescence, 2 gave borderline reac-
EVALUATION OF FLUORETEC FA SYSTEM
VOL. 10, 1979
TABLE 1. Reaction of 58 stock culture strains of the B. fragilis group with the Fluoretec BFG reagent. Fluorescence Bacteria
B. fragilis B. thetaiotaomicron B. vulgatus B. distasonis B. ovatus
No.
tested
Positive
Borderline
Negative
28 12 10 7 1
28 12 4 0 0
0 0 2 1 0
0 0 4 6 1
tions, and 4 failed to fluorescence. The 7 B. distasonis strains and the single B. ovatus strain failed to gave positive reactions. Smears of 27 blood cultures and 49 specimens of purulent material were stained by using the Fluoretec reagents. Of these 76 specimens, 45 were positive for members of the BFG by culture. Forty-four (98%) of the 45 culture-positive specimens were positive when stained with the Fluoretec BFG reagent. The 27 blood cultures showed complete agreement between FA and culture results (Table 2). The Fluoretec BFG reagent was positive for members of the BFG in 16 of the blood cultures, and in each case a member of the BFG was the only organism isolated. Of the remaining 11 blood cultures, which were negative for the BFG by FA and culture, one yielded a Clostridium and 10 yielded aerobic gram-negative bacilli. The one false-negative and the two false-positive results which were obtained from the 76 clinical specimens were from specimens of purulent material (Table 3). The single false-negative specimen (R78091) failed to give positive fluorescence even though B. fragilis was isolated. The direct smear of this specimen showed an abundance of brightly fluorescing material which appeared to be debris, but no bacteria with typical morphology and peripheral staining. The B. fragilis isolated from this specimen gave negative to weak overall fluorescence with an occasional cell staining more intensely. Two specimens of purulent material (R77008 and X78061) gave positive fluorescence for members of the BFG, but failed to yield these organisms by culture. One anaerobic gram-negative bacillus isolated from specimen X78061 was not able to be subcultured and therefore could not be identified or tested in pure culture with the FA reagents. The only other anaerobe isolated from this specimen failed to fluoresce with the BFG reagent as did the isolates from specimen R77008. Members of the BMG were detected by culture in 15 of the 76 clinical specimens. The Fluoretec BMG reagent detected members of the BMG in each of the 15 specimens (100%).
123
TABLE 2. Culture and FA results obtained from 27 blood cultures No. of cultures
16 1
FA
Culture
Isolates
BMG
BFG
FA
Culture
Aerobic
Anaerobic
-
None None
BFG Clostridium
+
+
-
-
-
-
sp. GNR' 10 a GNR, Gram-negative rod.
None
Five specimens were FA positive for the BMG, although these anaerobes could not be isolated. All aerobic and anaerobic isolates from each of the five specimens (X77081, X77082, R77084, R77098, and X78064) which demonstrated a false-positive reaction were tested individually with the BMG reagent. None of the organisms isolated from these clinical specimens gave positive results. The BFG and BMG reagents did not appear to react with aerobic or other anaerobic bacteria in clinical specimens; however, the BMG staining reagent did occasionally react weakly with Formalin-fixed BFG control cells. DISCUSSION The results obtained by testing 58 stock culture strains of the BFG show that the Fluoretec BFG reagent detected 100% of the 40 B. fragilis and B. thetaiotaomicron strains, but detected only 22% of 18 B. vulgatus, B. distasonis, and B. ovatus strains. The low percentage of detection of the latter three species did not appear to be a problem when testing direct clinical smears since there were no false-negative results due to the failure of these species to fluoresce. However, there were only two specimens from which these species were isolated in the absence of B. fragilis or B. thetaiotaomicron. The B. distasonis and the B. vulgatus strains present in these specimens did fluoresce, but two specimens is an insufficient number on which to evaluate the effectiveness of the BFG reagent for detection of B. vulgatus, B. distasonis, or B. ovatus in direct clinical smears. It should also be noted that results based on the data obtained with stock cultures (Table 1) may be misleading due to the small number of B. vulgatus, B. distasonis, and B. ovatus stock cultures tested and since some strains, after storage at -60°C, may not fluoresce with the same intensity as strains found in fresh clinical specimens. Although studies using stock cultures indicate inadequate coverage of these members of the BFG, the need for more complete FA coverage of B. vulgatus, B. distasonis, and B. ovatus may be minimized by the fact that these species are
124
J. CLIN. MICROBIOL.
HOLLAND, STAUFFER, AND ALTEMEIER
TABLE 3. Culture and FA results obtained from 49 clinical specimens ofpurulent material Culture results'
FA results
Source and culture no.
Abdominal fluid (R75086) Brain abscess (R77008) Liver abscess (R77023)
(R77024) Peritoneal fluid (X77062) Subphrenic abscess (R77064)
BFG
BMG
+
+
Aerobic
None B. asaccharolyticus Bacteroides sp. GPC
GNR GNR None GNR GPC
None None B. fragilis B. fragilis Bacteroides sp. F. nucleatum B. fragilis B. vulgatus B. thetaiotaomicron GPC None None None B. fragilis B. asaccharolyticus GPC (3) B. fragilis B. thetaiotaomicron B. distasonis B. vulgatus GPR None B. thetaiotaomicron GPC B. melaninogenicus GPR, GPC (2) Fusobacterium sp. Bacteroides sp. GPR GPC
Granulation tissue (R77066)
GNR (2) GPC GPR (2)
Subcutaneous abscess (R77073) Leg ulcer (R77077) Sinus tract (R77079) Gangrenous tissue (R77080)
None None GPR GNR (2) GPC (3)
Colon suture site (X77081)
GNR
Sinus tract (X77082) Osteomyelitis (R77084)
None GNR (3)
Pus (X77096)
None
Pleural fluid (X77097)
None
Fistula tract (R77098)
Colostomy drainage (X77099)
GNR GPC GNR (3) GPC
Osteomyelitis (X77101)
GNR GPC
Retroperitoneal abscess (R78001) Abdominal abscess
GNR
(R78002) (X78003)
GNR (2) GPC (2) GNR (2) GPC (2)
Pelvic abscess (X78004)
GPC
Buttock abscess (X78006)
GPC (2) GNR Yeast GPC (2) GNR (2) GNR (3) GNR
Jejunal wall (X78007) Peritoneal fluid (X78008) Subphrenic abscess (X78009)
Anaerobic
GNR None
B. fragilis B. melaninogenicus GPR, GPC (2) B. fragilis B. melaninogenicus GNR, GPC B. fragilis B. fragilis B. thetaiotaomicron B. thetaiotaomicron B. vulgatus GPC GPR B. fragilis GPC None GPR (2) GPR (2) B. fragilis
VOL. 10, 1979
EVALUATION OF FLUORETEC FA SYSTEM
125
TABLE 3-Continued Culture resultsa
FA results Source and culture no. BFG
BMG
Pilonidal abscess (X78011)
+
+
None
Sinus tract (X78012) Abdominal wound (R78018)
+
+
None GNR (2) GPC (2)
Pus, chest wall (R78021)
+
+
GPC (2) GNR
Pus (X78023)
+
+
GNR (2)
(X78024)
+
+
GNR
Buttock abscess (X78057)
+
+
ND
Bartholin abscess (X78061)
+
+
ND
Pilonidal cyst (R78062)
+
-
GPC
Fistula tract (R78063)
+
-
Abdominal wound (X78064) (X78067)
GNR GPC (4)
+ +
+ +
GNR GNR (2)
Lesion, groin (X78068)
+
+
GNR (2) GPC (2)
Amniotic fluid (X78069) Hip abscess (X78079) Abdominal wound (X78080) Facial abscess (X78086)
-
+
None GPC None None
+ + +
Aerobic
Anaerobic
B. fragilis B. asaccharolyticus GPR, GPC (3) None B. fragilis B. thetaiotaomicron B. distasonis B. melaninogenicus GPR (2), GPC (2) B. fragilis B. melaninogenicus GPC, GPR (2) B. fragilis B. asaccharolyticus GPC B. fragilis B. asaccharolyticus B. vulgatus B. melaninogenicus B. asaccharolyticus GNR B. thetaiotaomicron Bacteroides sp. B. fragilis GPC B. fragilis B. fragilis B. melaninogenicus Bacteroides sp. B. fragilis B. thetaiotaomicron B. melaninogenicus Bacteroides sp. None B. fragilis B. fragilis B. fragilis B. melaninogenicus GPC (3) None
Peritoneal exudate (X78088) None Sinus tract (R78090) GPC None (R78091) None B. fragilis Osteomyelitis (M818) GPC GPR Thoracentesis (M827) + None B. thetaiotaomicron a GNR, Gram-negative rod; GPR, gram-positive rod; GPC, gram-positive coccus; ND, not done. Numbers in parentheses indicate the number of isolates.
isolated from clinical specimens far less frequently than B. fragilis and B. thetaiotaomicron (7, 8), and when found, they are usually accompanied by one or both of these species. The Fluoretec BFG reagent was 100% accurate for detecting B. fragilis, B. thetaiotaomicron, and B. distasonis strains in blood cultures which contained gram-negative bacilli. The
aerobic gram-negative bacilli present in 10 of the blood cultures and the Clostridium species isolated from one specimen did not fluoresce, thus there were no false-positive reactions resulting from these bacteria. Based on this limited number of 27 specimens, it appears that the BFG reagent is a reliable tool for the detection of members of the BFG in blood cultures.
126
HOLLAND, STAUFFER, AND ALTEMEIER
There were no members of the BMG detected in the 27 blood cultures by either the BMG reagent or by culture; therefore, it was not possible to determine from the present study whether these anaerobes could be detected in blood cultures. Based on our results with specimens of purulent material (Table 3), there is reason to believe that a high percentage would be detected; however, the reported incidence of these anaerobes in blood cultures (3, 10) is probably not high enough to justify routine testing for the BMG. With direct smears of purulent material, 98% of the specimens which were positive for members of the BFG by culture were also positive with the Fluoretec BFG reagent. These results are similar to those obtained from blood cultures; however, direct smears of some purulent materials were more difficult to evaluate than blood culture specimens because they contained greater amounts of debris and fewer bacterial cells with typical morphology. These factors probably influenced the evaluation of specimen R78091 (Table 3). In addition, blood cultures were not stained unless a gram-negative bacillus was detected by Gram stain or subculture. Therefore, blood culture specimens were less likely to yield false-positive results due to factors such as prior antibiotic treatment and inadequate care in collection, transportation, and cultivation which could influence the results obtained from specimens that were stained before the demonstration of viable bacteria. These factors may have been responsible for the falsepositive results obtained from specimens R77008 and X78061 (Table 3). False-positive results could also have resulted from the failure to recognize and pick all of the different colony types present on the original isolation plates. Five clinical specimens were positive for members of the BMG when stained with the Fluoretec BMG reagent, yet these bacteria were not detected by culture. This could have resulted from the reasons previously discussed with regard to the BFG. Occasionally, a small percentage of Formalin-fixed BFG control cells fluoresced weakly when reacted with the BMG reagent. This was probably not the cause of the five false-positive reactions since one specimen yielded no growth and the isolates from the remaining four specimens did not fluoresce with the BMG reagent. During this study, the Fluoretec reagents were reliable tools for detecting members of the BFG and the BMG in clinical specimens. The real advantage of using the FA procedure is that members of the BFG and the BMG can be detected in purulent material within 1 to 2 h after the specimen is obtained and in blood
J. CLIN. MICROBIOL.
cultures as soon as or possibly before growth is detected. Also, since bacterial viability is not required, detection may be possible even if the specimen is improperly handled during collection and transportation or if the specimen is obtained after the administration of certain antibiotics. The FA procedure should improve patient care by reducing by several days the time required to presumptively detect the BFG and the BMG, thus allowing the early initiation of
appropriate therapy. As with any FA procedure, the interpretation of results will depend a great deal on the experience and judgment of the individual performing the test, the type and condition of equipment, the proper use of test and control reagents,
and the criteria used to determine what constitutes a positive and a negative result. For this study, clinical specimens were judged to be positive if bacterial cells showed peripheral fluorescence and negative if peripheral fluorescence was not demonstrated. Most specimens gave clear-cut results; however, some gave borderline reactions which were difficult to evaluate. Several factors were found to influence the amount of fluorescence detected. The intensity of fluorescent staining could frequently be increased by extending to 30 min the reaction period of the FA reagent with the specimen. With some specimens, fluorescence was difficult to detect because cells did not adhere sufficiently to slides after heat fixation and were lost during the rinsing and blotting steps. In our experience, a 10-min methanol fixation of airdried slides resulted in greater cell adherence. The type of excitation filter used also affected the degree of staining. Reactions which were questionable with a BG 12 filter were frequently positive with an FITC interference type filter. The proper alignment and use of the microscope are other factors to be considered. It is recommended that potential users of the FA technique be given training in fluorescent microscopy which would include the proper use and care of equipment, the proper use of the test and control reagents, and the opportunity to evaluate known specimens. ACKNOWLEDGMENTS We thank H. S. Bjornson for a critical review of the manuscript. This work was supported in part by Public Health Service grant 5-P50-GM-15428 from the National Institute of General Medical Sciences and in part by Pfizer Inc., New York, N.Y.
LITERATURE CITED 1. Cato, E. P., and J. L. Johnson. 1976. Reinstatement of species rank for Bacteroides fragilis, B. ovatus, B. distasonis, B. thetaiotaomicron, and B. tulgatus: designation of neotype strains for Bacteroides fragilis (Veillon and Zuber) Castellani and Chalmers and Bac-
VOL. 10, 1979
EVALUATION OF FLUORETEC FA SYSTEM
teroides thetaiotaomicron (Distaso) Castellani and Chalmers. Int. J. Syst. Bacteriol. 26:230-237. 2. Fales, W. H., and G. W. Teresa. 1972. Fluorescent antibody technique for identifying isolates of Sphaerophorus necrophorus of bovine hepatic abscess origin. Am. J. Vet. Res. 33:2323-2329. 3. Felner, J. M., and V. R. Dowell, Jr. 1971. "Bacteroides" bacteremia. Am. J. Med. 50:787-796. 4. Finegold, S. M., and E. A. Barnes. 1977. Report of the ICSB taxonomic subcommittee on gram-negative anaerobic rods: proposal that the saccharolytic and asaccharolytic strains at present classified in the species Bacteroides melaninogenicus (Oliver and Wherry) be reclassified in two species as Bacteroides melaninogenicus and Bacteroides asaccharolyticus. Int. J. Syst. Bacteriol. 27:388-391. 5. Garcia, M. M., D. H. Neil, and K. A. McKay. 1971. Application of immunofluorescence to studies on the ecology of Sphaerophorus necrophorus. Appl. Microbiol. 21:809-814. 6. Griffin, M. H. 1970. Fluorescent antibody techniques in
127
the identification of the gram-negative nonsporeforming anaerobes. Health Lab. Sci. 7:78-83. 7. Holland, J. W., E. 0. Hill, and W. A. Altemeier. 1977. Numbers and types of anaerobic bacteria isolated from clinical specimens since 1960. J. Clin. Microbiol. 5:2025.
8. Holland, J. W., S. M. Gagnet, S. A. Lewis, and L. R. Stauffer. 1977. Clinical evaluation of a simple, rapid procedure for the presumptive identification of anaerobic bacteria. J. Clin. Microbiol. 5:416-426. 9. Lambe, D. W., Jr. 1974. Determination of Bacteroides melaninogenicus serogroups by fluorescent antibody staining. Appl. Microbiol. 28:561-567. 10. Martin, W. J. 1974. Isolation and identification of anaerobic bacteria in the clinical laboratory: a 2-year experience. Mayo Clin. Proc. 49:300-308. 11. Stauffer, L. R., E. 0. Hill, J. W. Holland, and W. A. Altemeier. 1975. Indirect fluorescent antibody procedure for the rapid detection and identification of Bacteroides and Fusobacterium in clinical specimens. J. Clin. Microbiol. 2:337-344.
